E
Elon Portugaly
Researcher at Microsoft
Publications - 5
Citations - 751
Elon Portugaly is an academic researcher from Microsoft. The author has contributed to research in topics: Counterfactual thinking & Leverage (statistics). The author has an hindex of 4, co-authored 5 publications receiving 628 citations.
Papers
More filters
Journal Article
Counterfactual reasoning and learning systems: the example of computational advertising
Léon Bottou,Jonas Peters,Joaquin Quiñonero-Candela,Denis X. Charles,D. Max Chickering,Elon Portugaly,Dipankar Ray,Patrice Y. Simard,Ed Snelson +8 more
TL;DR: This work shows how to leverage causal inference to understand the behavior of complex learning systems interacting with their environment and predict the consequences of changes to the system and allow both humans and algorithms to select the changes that would have improved the system performance.
Journal ArticleDOI
Counterfactual Reasoning and Learning Systems
Léon Bottou,Jonas Peters,Joaquin Quiñonero-Candela,Denis X. Charles,D. Max Chickering,Elon Portugaly,Dipankar Ray,Patrice Y. Simard,Ed Snelson +8 more
TL;DR: In this article, the authors leverage causal inference to understand the behavior of complex learning systems interacting with their environment and predict the consequences of changes to the system, allowing both humans and algorithms to select the changes that would have improved the system performance.
Posted Content
Counterfactual Reasoning and Learning Systems
Léon Bottou,Jonas Peters,Joaquin Quiñonero-Candela,Denis X. Charles,D. Max Chickering,Elon Portugaly,Dipankar Ray,Patrice Y. Simard,Ed Snelson +8 more
TL;DR: This work shows how to leverage causal inference to understand the behavior of complex learning systems interacting with their environment and predict the consequences of changes to the system and allow both humans and algorithms to select changes that improve both the short-term and long-term performance of such systems.
Patent
Considering user-relevant criteria when serving advertisements
TL;DR: In this article, a criterion (e.g., keyword, image, audio element, etc.) is selected to evoke the advertisement, regardless of whether the criterion appears in the content of the webpage, and the expected gain of presenting the criterion-evoked advertisement.
Posted Content
Unbiased Estimation of the Value of an Optimized Policy.
TL;DR: This work presents a procedure for learning optimized policies and getting unbiased estimates for the value of deploying them, and proves that inverse-propensity-weighting effect estimator is unbiased when applied to the optimized subset.